Fig. 3.  The association of the outpatient choice of β-blockade with postoperative mortality (Cox survival analysis log-rank test P  value, P < 0.0001) based on the outpatient preoperative β-blocker choice is demonstrated. The population included 1,871 patient who were on a β-blocker as an outpatient and remained on that specific β-blocker through their inpatient hospital stay.